Динамическое программирование и планирование сценариев в задаче оптимизации перевозок контейнерных грузов
Within the last twenty years methods of optimization simulation began to be used actively in planning and supply chain management. Considering specifics of problems of planning in supply chains the greatest popularity received the linear programming and its such methods as dynamic programming, stochastic programming and planning of scenarios. These methods allow to perform optimization of the supply chain according to numerous databases, to each of which there corresponds the scenario describing different options of development in the uncertain future.
Despite rather intensive researches in this area, stochastic programming is still poorly developed section of mathematical programming. In stochastic programming it is more, than in other sections of the theory of the conditional extremum problems, the considerable difficulties arise not only by development of solution methods, but also in case of problem definition in which it is necessary to reflect sometimes quite thin situations of planning and control in the conditions of risk and uncertainty. From there is the need for development of new models of planning and supply chain management in the conditions of the incomplete information and methods allowing them to research.
In paper considers the problem of application of method of dynamic linear programming for the solution of the task of the choice of optimum strategy of sending containers taking into account essential irregularity of loading and sending containers is considered. The mathematical model of the multistage dynamic problem about sending loaded containers which allows to perform the choice of optimum strategy of sending containers taking into account essential irregularity of their loading and sending from railway freight yard is offered. The numerical example which is set by efficiency of use of the method of dynamic linear programming for the solution of the considered task is provided.